基于时序分析的网络业务分析与预报  

Network Traffic Modeling and Forecasting Based on Time Series Analysis

作　　者：薄今纲[1] 刘嘉火昆 金志刚[1] 余晖[2] 

机构地区：[1]天津大学,天津300072 [2]天津移动通信公司,天津300021

出　　处：《计算机应用研究》2003年第9期26-28,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(90104015);IBMSUR;刘徽应用数学中心资助项目(T08);天津移动通信公司资助项目

摘　　要：随着网络应用的迅速发展,提供保证一定质量的通信服务(QoS)成为一个重要的研究方向,而网络业务的分析与预报是保证QoS的关键技术。分析了天津市163,169,2163的拨号上网业务在线总时长,应用时间序列分析的方法对网络业务进行分析,并在模型基础上进行准确的预报。其结果对网络容量进行规划并提前进行相应的调整或升级,确保网络的整体运行状况是有意义的。With the great development of Internet utility,the kinds of Internet services have become more and more.So it has been an important research orientation to bring the communication service of great quantity,however,the analysis and forecast of Internet business is the key of it.In this paper,we analyze and handle total online time of the users for 163,169,2163 in Tianjin.We use the basic theory of time series to model with the traffic and then forecast on the model.The result can regulate the capacity of Internet and adjust or promote correspondingly,thereby,we can guarantee the overall operation condition of Internet.

关 键 词：网络业务 ARIMA模型 预报 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]